Sensitive UV photoacoustic detection of ozone

Sensitive UV photoacoustic detection of ozone Tse-Luen Wee and A. H. Kung* Institute of Atomic & Molecular Sciences, Academia Sinica, Taipei 106, Taiwan R.O.C. Tel. 886-2-23668270, Fax 886-2-23620200 E-mail:[email protected]

Abstract We describe the detection of trace amount of ozone in a gas mixture at atmospheric pressure using excitation by the fourth harmonic of a Nd:YAG laser at 266 nm and measurement by photoacoustic detection. We show that a sensitivity of 10 ppbV can be achieved.

1. Introduction: Ozone is a strong oxidizing agent. In the earth’s upper atmosphere, ozone occurs naturally and protects us by shielding the sun’s harmful rays. In the lower atmosphere, ozone is formed when pollutants react chemically in the presence of sunlight. [1] A controlled and suitable amount of ozone in the environment can be used to kill germs and bacteria in air and water, and to remove toxic chemicals from fruits and vegetables. However, too much ozone can be harmful. Prolonged exposure to a high concentration of ozone will affect the respiratory functions of the human lung and other organs and cause significant damage to agriculture, physical properties and the environment. Ozone exposure has been linked with adverse effects. Nose and throat irritation, respiratory symptoms and decreases in lung function have been observed in healthy, exercising persons breathing air containing elevated levels of ozone. Respiratory symptoms include shortness of breath, chest pain and coughing, and may occur in both adults and children. Animal studies have shown that ozone damages sensitive lung tissue and effects may continue for some time after exposure has ended. [2] The major source of atmospheric ozone is a result of human activities in using fossil fuel products.

The major constituents of smog, oxides of nitrogen and volatile hydrocarbon

compounds, react with oxygen under sunlight to form ozone.

150

Nonlinear Frequency Generation and Conversion: Materials, Devices, and Applications III, edited by Kenneth L. Schepler, Dennis D. Lowenthal, Proceedings of SPIE Vol. 5337 (SPIE, Bellingham, WA, 2004) · 0277-786X/04/$15 · doi: 10.1117/12.528906

Downloaded from SPIE Digital Library on 16 Jan 2012 to 140.114.195.186. Terms of Use: http://spiedl.org/terms

With the increasing use of ozone for sterilization applications in large facilities such as hospitals and factories and in metropolitan areas, it is essential to have accurate and reliable monitoring of the amount of ozone in the environment. The safe level of ozone is established at 0.1 parts per million. Some countries set a safe level at 0.05 ppmV for exposure in an eighthour period. Because of the large health and economic impact due to exposure to ozone, it is necessary to have sensitive, accurate, reliable detectors of ozone. There are two types of ozone detectors that are in common use. [3] The first method relies on chemiluminescence that results from ozone reacting with another chemical. This older approach is not very stable and the material has a short useful life. The second method is by the use of photometry to measure the reduction of UV quanta due to absorption by ozone at 254 nm. The light source used is the HgI emission line at 253.7 nm. This is the most commonly used approach to date. However, the method has significant shortcomings that limit its utility in the commercial market. These shortcomings result from insufficient sensitivity of the photometric approach which requires an extremely stable light source and complex multiple pass optical arrangement to reach a sub-ppmV level of sensitivity. Other types of detectors are less advanced or too expensive to be commonly used at this point.

2. Photoacoustic detection A successful application of the photoacoustic effect has been for trace detection of gases. Trace gas sensing systems have to meet several requirements like high detection sensitivity and selectivity, multicomponent capability, field suitability, etc. Today, laser-based gas sensing devices with modern laser technology and photoacoustic detection schemes are versatile and powerful instruments for all kinds of applications. In comparison to competing optical detection schemes, photoacoustics offers the distinct advantage of a rather simple and robust setup, room temperature operation, wavelength independent detection, and compact size. [4] The photoacoustic technique is a method of directly detecting light absorption as opposed to photometry. It has been shown that the acoustic wave signal amplitude is linearly proportional to the amount of light absorbed. Since microphones are very sensitive and have a large dynamic range, the photoacoustic technique is widely used to detect trace amount of molecules in a dense Proc. of SPIE Vol. 5337


151

medium with high accuracy. The technique has achieved a sensitivity that is often better than one part per billion. More than 50 different gases have been shown to achieve this kind of sensitivity using photoacoustic detection. The dynamic range of the detected gas concentration can easily be over one million. [5] The selectivity of the photoacoustic effect, where a narrowband laser excites only those species with absorptions at the laser wavelength, is also a valuable feature of the photoacoustic effect in its application to trace detection. Current photoacoustic trace gas detection techniques use almost exclusively an IR source which provides high selectivity as well as sensitivity. The IR absorption of ozone has a near coincidence with the 9P (14) emission line of the CO2 laser. The CO2 laser has thus been used for photoacoustic measurement of small quantities of ozone in a gas mixture. [6, 7] However, since the ozone absorption coefficient in the IR is weak (41.7 x 10-20 cm2), a high power laser together with a multiple pass absorption scheme is required. The apparatus is complex and not easy to operate in the field. Nevertheless, a sensitivity of 10 ppbV is reported. [6]

Figure 1. Absorption spectrum of ozone

152

Proc. of SPIE Vol. 5337


On the other hand, it is well known that ozone has a strong and diffuse absorption in the uv centered around 260 nm. As shown in figure 1, this absorption band has a cross-section of 1 x 10–17 cm2. This is a factor of 24.6 larger than in the IR. Radiation from the commonly used lasers of the KrF excimer laser at 248 nm or the fourth harmonic of the Nd:YAG laser at 266 nm will be absorbed strongly by ozone. We report here our preliminary results of using the fourth harmonic of the Nd:YAG laser for photoacoustic detection of ozone in a gas mixture.

3. Experimental setup: The UV photoacoustic ozone detector setup used in our experiment was consisted of four major parts. Part one was a 266 nm laser source. The second was a resonant acoustic cell. The third was an ozone generator, and the fourth was an electronic signal processing system. This is shown in figure 2. We used a Nd-YAG laser at 4 kilohertz repetition rate, making it possible to match the acoustic resonance frequency of our resonant photoacoustic cell. A KTP crystal and a BBO crystal were used to generate the fourth harmonic at 266 nm. The output uv power was 20 mW. A fused silica lens was used to focus the uv beam through the photoacoustic cell. We measured acoustic signals generated by passing a controlled amount of ozone through the cell. Ozone was obtained from a commercial domestic ozone generator which generates ozone by pumping room air through a corona discharge. The amount of ozone thus generated can be varied up to 200 milligram per hour with plus minus 10% in accuracy.



153

Lock-in amplifier (PC-controlled) Acoustic amplifier

1064nm 532nm

Photoacoustic cell

KTP

1064nm

266nm

Nd:YAG laser BBO Power-meter

Ozone in Ozone out

Ozone Generator

Figure 2. UV photoacoustic ozone scheme.

The resonant photoacoustic detector employed was specifically designed for fast time response, low acoustic and electrical noise characteristics, and high sensitivity. The acoustic cell design followed that first used by Andras Miklos.[8] To reduce the flow noise and external electromagnetic disturbances from electronic devices two symmetric cylindrical acoustic resonators (5.5 mm diameter tubes) were placed in parallel in the PA cell, two quarter-wave filters on both ends of the resonator and additional filters to reduce gas flow noise. This design minimizes background from the windows. Differential signal processing balanced off external acoustic interference. It allowed continuous gas flow and can be used in a noisy environment. For a ~4 kHz resonance frequency, the resonator tube diameter was 8 mm and the length was 40 mm.

The dimensions of the filter were 20 mm cubed.

A pair of matched microphones

(Knowles 3029) was placed in the central section of each tube to receive the photoacoustic signal and an amplifier amplified the differential signal from the microphones.

154



3. Results and discussion: Figure 3 shows the results of the experiment.

The lower trace is a measure of the

transmitted 266 nm power as the ozone concentration of the gas mixture that flows through the photoacoustic cell was changed as a function of time. The amount of ozone was increased in a stepwise manner with time by adjusting the current in the corona discharge except for the last time segment where the ozone generator current was turned off. The corresponding ozone concentration varied from 30 ppmV to 120 ppmV. Due to the large absorption cross-section of ozone at 266 nm, the amount of uv power absorbed was as much as 55% at the highest concentration used in the experiment. The associated photoacoustic signal is shown in the upper trace. As could be expected, the photoacoustic signal increased with the ozone concentration.

Careful comparison of the

photoacoustic signal and the uv power indicates that the photoacoustic signal is linear with the absorbed power as predicted by theory. The maximum signal from the lock-in amplifier reached 350 mV. The ripples seen on each time segment was due to slow periodic fluctuation in the uv power, possibly due to thermal effect on the BBO crystal. The acoustic amplifier gain under the conditions of this measurement was equal to four. The background signal caused by window

PA signal (mV)

absorption was